Simulations

Simulations are complete training sessions that bundle one or more scenarios together with rubrics, time limits, and ordering — they are what learners actually launch to begin practicing.

What is a Simulation?

A simulation is a runnable training package. It wraps one or more Scenarios and adds session-level configuration: which rubrics to evaluate against, how much time the learner has per scenario, the order scenarios appear, and behavioral flags.

A simulation contains:

• Name — a descriptive title (e.g., “Academic Integrity Training”, “Aggressive Practice”)
• Description — what the simulation covers and its purpose
• Scenarios — the training situations included in this simulation, each with per-scenario configuration:
  • Rubrics — evaluation criteria applied to that scenario (e.g., Communication Skills, Policy Knowledge, De-escalation)
  • Flags — per-scenario boolean settings
  • Time Limits — maximum duration for the learner on that scenario
  • Positions — the order in which scenarios are presented
• Departments — organizational groupings for filtering and access control
• Flags — simulation-level boolean settings

Simulations are assigned to Cohorts, which control which learners have access. When a learner starts a simulation, Glow creates an Attempt with one Chat per scenario.

How Simulations Connect to the Workflow

Simulations are step 3 in the Glow content pipeline:

Create a simulation

Via the CLI

Create an “Academic Integrity Training” simulation:

Calls below use $GLOW_INSTANCE_URL + $GLOW_TOKEN — see Authentication to export them once.

glow simulations create --body '{
  "simulations": [{
    "name": "Academic Integrity Training",
    "description": "Structured training for handling academic integrity violations"
  }]
}'

Via the API

curl -X POST $GLOW_INSTANCE_URL/simulation/create \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -d '{
    "simulations": [{
      "name": "Academic Integrity Training",
      "description": "Structured training for handling academic integrity violations"
    }]
  }'

Assigning scenarios and per-scenario configuration

Each scenario added to a simulation gets its own per-scenario configuration: rubrics, time limits, positions, and flags. This is managed through the simulation draft endpoint.

glow simulations draft --body '{
  "input_draft_id": "simulation-draft-uuid",
  "scenario_ids": ["academic-integrity-scenario-uuid"],
  "scenario_rubrics": [
    {"scenario_id": "academic-integrity-scenario-uuid", "rubric_id": "communication-skills-rubric-uuid"},
    {"scenario_id": "academic-integrity-scenario-uuid", "rubric_id": "policy-knowledge-rubric-uuid"}
  ],
  "scenario_time_limits": [
    {"scenario_id": "academic-integrity-scenario-uuid", "time_limit": 600}
  ]
}'

Rubrics

Rubrics define how the learner’s performance is evaluated for each scenario. Available rubrics in the university seed data include:

• Communication Skills — evaluates clarity, empathy, and professionalism
• Policy Knowledge — evaluates correct application of relevant policies
• De-escalation — evaluates ability to reduce tension in confrontational situations

Assign rubrics per scenario:

glow simulations draft --body '{
  "input_draft_id": "simulation-draft-uuid",
  "scenario_rubrics": [
    {"scenario_id": "scenario-uuid-1", "rubric_id": "communication-skills-uuid"},
    {"scenario_id": "scenario-uuid-1", "rubric_id": "de-escalation-uuid"},
    {"scenario_id": "scenario-uuid-2", "rubric_id": "policy-knowledge-uuid"}
  ]
}'

Time Limits

Set maximum durations per scenario to simulate real-world time pressure:

glow simulations draft --body '{
  "input_draft_id": "simulation-draft-uuid",
  "scenario_time_limits": [
    {"scenario_id": "scenario-uuid-1", "time_limit": 600},
    {"scenario_id": "scenario-uuid-2", "time_limit": 300}
  ]
}'

Positions

Control the order in which scenarios are presented to the learner:

glow simulations draft --body '{
  "input_draft_id": "simulation-draft-uuid",
  "scenario_positions": [
    {"scenario_id": "scenario-uuid-1", "position": 0},
    {"scenario_id": "scenario-uuid-2", "position": 1}
  ]
}'

Simulation Patterns

The university seed data demonstrates two common simulation patterns:

Structured Training Simulations

Multi-scenario simulations designed to train a specific skill. Scenarios are presented in a fixed order and each has rubrics attached:

• Academic Integrity Training — a complete training session on handling cheating incidents
• FERPA Training — end-to-end practice with FERPA violation procedures
• Upset Student Training — structured practice for de-escalation in office hours

Open Practice Simulations

Single-scenario simulations designed for free-form practice with a specific personality type. These are lower-stakes and let the learner experiment:

• Confused Practice — practice with a student who asks clarifying questions
• Happy Practice — practice with a positive, uplifting student
• Passive Practice — practice with a disengaged student
• Aggressive Practice — practice with a confrontational student
• General Practice — open-ended interaction practice

Working with Drafts

Simulations use the draft system for all edits. The draft endpoint supports optimistic concurrency via expected_version.

# Create or update a draft
glow simulations draft --body '{
  "input_draft_id": "existing-draft-uuid",
  "expected_version": 2,
  "name": "Updated Academic Integrity Training",
  "description": "Revised training with new de-escalation scenarios",
  "scenario_ids": ["scenario-uuid-1", "scenario-uuid-2"]
}'

# Via API
curl -X PATCH $GLOW_INSTANCE_URL/simulation/draft \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -d '{
    "input_draft_id": "existing-draft-uuid",
    "expected_version": 2,
    "name": "Updated Academic Integrity Training"
  }'

List your drafts:

# CLI
glow simulations list
 
# API
curl -X POST $GLOW_INSTANCE_URL/simulation/drafts \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GLOW_TOKEN"

Search and filter

Simulation search supports full-text plus facet filters across scenario, department, and flag.

glow simulations search --body '{
  "search": "training",
  "filter_department_ids": ["dept-cs"],
  "page_size": 25
}'

curl -X POST $GLOW_INSTANCE_URL/simulation/search \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"search": "training", "page_size": 25}'

Bulk operations

Bulk delete and update follow the canonical all-matching shape — pass either explicit IDs, or all: true with flat filter fields + optional excluded_ids + a patch for updates.

# Delete every simulation in a sunset department, except a manual keeplist
glow simulations delete --body '{
  "all": true,
  "filter_department_ids": ["dept-sunset"],
  "excluded_ids": ["sim-keep-1", "sim-keep-2"]
}'
 
# Bulk update via patch
glow simulations update --body '{
  "all": true,
  "filter_department_ids": ["dept-archive"],
  "patch": { "archived": true }
}'

See the Personas guide for the canonical write-up of this pattern.

Common Operations

Related

• Simulations API Reference — full endpoint schemas and field definitions
• Simulations CLI Reference — all CLI commands and flags
• Scenarios Guide — creating the training situations that simulations contain
• Personas Guide — the AI characters used within scenarios
• Attempts Guide — what happens when a learner starts a simulation
• Agents Guide — configuring the AI engine that powers simulations